A Review of Advances and Challenges in Intelligent Disaster Assessment of High-Value Objects in Remote Sensing Imagery
-
摘要: 随着全球自然灾害频发以及各类突发事件风险上升，如何依托智能遥感技术对受灾目标开展快速、精准的评估，已成为支撑应急响应与灾后重建的重要任务。近年来，基于深度学习的遥感灾害评估方法不断涌现，大大提高了灾害评估的精度和智能化程度，然而，该领域发展迅速但仍缺乏系统性综述。本文全面梳理了遥感目标灾害智能评估的技术内涵；对现有研究中常用的目标损伤等级标准、公开数据集及性能评价指标进行了归纳；围绕双时相变化检测、多时相序列建模、多模态数据融合、数据受限场景下灾害评估四类技术框架进行深入分析，叙述了各类方法的技术路径及其优势与不足；最后，本文进一步探讨了该领域未来的发展方向，旨在为遥感智能灾害评估技术在人为及自然灾害应对中的应用提供理论支持与方法参考。Abstract:
Significance The rapid and accurate disaster assessment of objects in the aftermath of disasters is critical for enabling effective emergency response and post-disaster recovery operations. Intelligent remote sensing technologies, empowered by deep learning, provide scalable, objective, and efficient means of assessing disaster across large and complex environments, such as densely populated urban centers, transportation hubs, and critical energy infrastructure. By leveraging high-resolution satellite and aerial imagery, these technologies can provide timely situational awareness for decision-makers, supporting prioritization in rescue and recovery tasks. Despite significant advancements in methods and applications, there remains a lack of comprehensive synthesis in the field, leading to fragmented practices and inconsistent benchmarks across studies. This study addresses this critical gap by providing a structured and systematic review that consolidates technical foundations, commonly used datasets, evaluation metrics, and methodological advances in deep learning-based remote sensing disaster assessment. The overarching goal is to accelerate the adoption and operationalization of these technologies in real-world disaster scenarios, ultimately contributing to the construction of resilient cities and infrastructures under increasing environmental and geopolitical risks. Progress In recent years, deep learning-based remote sensing disaster assessment methods have developed rapidly, demonstrating notable improvements in classification accuracy, processing automation, and scalability. Significant advances include bi-temporal change detection methods that can precisely localize disaster by capturing differences before and after disaster events, and multi-temporal sequence modeling approaches that extract evolving disaster patterns and degradation trends across time-series data. Furthermore, multi-modal data fusion strategies that combine optical, Synthetic Aperture Radar(SAR), and Light Detection and Ranging (LiDAR)data have enhanced the ability to analyze complex disaster characteristics under varying observation conditions. Advanced techniques to address data scarcity, such as transfer learning and self-supervised learning, have further extended the applicability of disaster assessment methods in data-constrained environments. These advances collectively contribute to improving the responsiveness and effectiveness of disaster assessment systems in supporting emergency decision-making and resource allocation during critical disaster events. Conclusions This study systematically categorizes and evaluates the technical landscape of deep learning-based remote sensing disaster assessment for objects, highlighting the strengths and limitations of bi-temporal, multi-temporal, multi-modal, and data-scarce scenario methods. While existing approaches demonstrate considerable potential in addressing various challenges in post-disaster assessment, challenges remain in ensuring robustness across diverse environments and operational conditions. The review underscores the need for standardized disaster classification criteria and comprehensive evaluation frameworks that consider both physical disaster and functional impacts to facilitate practical and consistent disaster assessments in real-world deployments. Prospects Future research in remote sensing-based disaster assessment should focus on developing multi-level collaborative frameworks for evaluating diverse objects across spatial and functional scales, enabling holistic disaster impact assessment that captures both direct and cascading effects. Complex scenarios such as airports, industrial zones, and ports contain static structures, moving objects, and interdependent functional units, thus requiring hierarchical modeling and multi-object reasoning. Physics-driven and hybrid approaches integrating structural mechanics, material degradation, and expert knowledge can further improve interpretability and generalization. Meanwhile, lightweight model design and edge deployment are important for real-time assessment on drones and satellites in emergency situations. Standardized evaluation metrics that combine physical disaster mechanisms with functional impact analysis will also be essential for practical deployment. Together, these directions will help transform intelligent remote sensing technologies into actionable tools for disaster response and recovery. -
Key words:
- Intelligent disaster assessment /
- Remote sensing image /
- Deep Learning
-
表 1 不同灾害类型对比
灾害类型 形成机制 典型灾害事件(2020～2025年) 影响范围 地震 地面震动、剪切波传播 2023年土耳其-叙利亚地震；
2025年缅甸地震城市建筑不同程度倒塌，人员伤亡，国际救援介入 飓风/台风 强风、暴雨、风暴潮 2020年美国飓风“劳拉”；
2023年台风“杜苏芮”登陆菲律宾/福建洪水淹城，通讯中断，经济损失巨大 洪水 降水过量、堤坝溃决 2021年河南郑州“7·20”特大暴雨；
2023年中国河北涿州洪灾地铁被淹、城镇基础设施受灾、
大规模房屋损坏和人员伤亡野火 干燥高温、引燃源 2023年加拿大野火；
2025年美国洛杉矶山火森林烧毁，森林周边居民区房屋受灾 海啸 海底地震引发海浪 2021年印尼马鲁古海小型海啸(局地)；
2022年汤加海底火山喷发引发海啸沿岸建筑受灾，通联中断，
邻国多地发出预警地区冲突
/爆炸爆炸、穿甲弹、燃烧弹 巴以冲突；俄乌冲突；伊以冲突 城市废墟化、基础设施毁坏、能源/粮食危机、
难民危机蔓延至多国
下载: 导出CSV
表 2 四级联合损伤等级标准
损伤等级 地震 飓风/台风 洪水 野火 海啸 爆炸 无损伤 未受影响。 轻度损伤 表面裂缝、非结构构件损伤，结构基本完整。 屋顶覆盖物(瓦片、金属板、防水层)局部损坏或掀起。 建筑底部或周边出现短期、浅层积水。 建筑外立面或屋顶局部烧蚀，主体结构保持稳定。 外立面或非承重构件受冲刷或撞击。 建筑局部出现破孔或外立面损伤。 重度损伤 明显结构性破坏，部分构件失效或局部坍塌。 屋顶大面积掀翻或结构性破坏，功能严重受限。 建筑长时间被洪水或泥水包围，发生明显侵蚀。 大面积烧毁，发生明显结构损伤。 承重墙、楼板或基础发生明显结构性破坏，功能基本丧失。 建筑出现大范围结构破坏，屋顶、楼板或承重墙部分坍塌。 完全损伤 建筑整体或近乎整体倒塌，结构完全失效。 主体结构不复存在，无法继续使用修复。 完全被积水/泥水覆盖，主体严重破坏。 建筑整体烧毁，仅残留地基或局部残骸。 建筑被整体冲毁、掀走或完全倒塌。 建筑整体倒塌、消失或不可识别。
下载: 导出CSV
表 3 常见数据集介绍
名称 灾害类型 图像
模态图像来源 标注类别 图像数
(张)图像尺寸/
分辨率(米)受灾地区 xBD[8]
(2019)地震/海啸、洪水、飓风、野火、火山喷发 卫星光学图像 Maxar数字地球开放数据计划 5类：背景、无损伤、轻度损伤、重度损伤、完全损伤 22068 1024 ×1024
/ 0.5-0.8美国、印度尼西亚、菲律宾等全球15个国家的受灾地区 Ida-BD[16]
(2022)飓风 卫星光学图像 WorldView-2卫星 5类：背景、无损伤、轻度损伤、重度损伤、完全损伤 174 1024 ×1024
/ 0.5美国路易斯安那州新奥尔良部分区域 RescueNet[17]
(2023)飓风 无人机光学图像 机器人辅助搜救中心(小型无人机系统) 10类：背景、水、无损伤建筑、中度损伤建筑、重损伤建筑、完全损伤建筑、车辆、树、水池、阻塞道路 4494 3000 ×4000
/ —美国佛罗里达州墨西哥海滩附近 QQB[18]
(2024)地震 卫星光学与SAR图像 WorldView-3卫星/Capella Space的GEO产品 2类(建筑物级别，不是像素级别)：无损伤、
有损伤16116 建筑物大小
/ SAR：0.35
光学：0.31土耳其、叙利亚 BRIGHT[19]
(2025)地震、洪水、海啸、火山喷发、野火5种灾害，地区冲突、爆炸2种人为灾害 卫星光学图像与SAR图像 Maxar数字地球开放数据计划/Capella Space雷达卫星/Umbra卫星 4类：背景、无损伤、
有损伤、完全损伤4246 1024 ×1024
/ 0.3-1土耳其、墨西哥、缅甸等全球14个国家的受灾地区 WCP[20]
(2023)地区冲突 卫星光学图像 哨兵-2号卫星、Google earth影像 3类：背景、
无损伤、有损伤81 120×120,
6×6, 3×3
/ 0.5, 106个叙利亚城市和4个乌克兰城市
下载: 导出CSV
表 4 常用评估指标
指标 计算公式 意义 准确率 (Accuracy) $ \mathrm{Accuracy}=\dfrac{\mathrm{TP}+\text{TN}}{\mathrm{TP}+\mathrm{TN}+\mathrm{FP}+\text{FN}} $
TP为真正例：被模型预测为正类的正样本
TN为真反例：被模型预测为负类的负样本
FP为假正例：被模型预测为正类的负样本
FN为假反例：被模型预测为负类的正样本衡量模型预测正确的比例，
反映灾害评估模型的总体能力精确率 (Precision) $ \mathrm{Precision}=\dfrac{\text{TP}}{\mathrm{TP}+\text{FP}} $ 在所有被模型预测为损伤的目标中，实际为损伤的比例，
强调“误报率”低召回率 (Recall) $ \mathrm{Recall}=\dfrac{\text{TP}}{\mathrm{TP}+\text{FN}} $ 在所有真实损伤目标中，被模型成功识别出来的比例，
反映“漏报率”低F1分数 $ \text{F1}=2⋅\dfrac{\text{Precision⋅Recall}}{\text{Precision+Recall}} $ 精确率与召回率的调和平均值，综合评估模型预测的
准确性与完整性交并比 (IoU) $ \mathrm{IoU}=\dfrac{\left| \mathrm{A}\cap \mathrm{B}\right| }{\left| \mathrm{A}\cup \mathrm{B}\right| } $
A：预测区域像素集合
B：真实区域像素集合损伤识别结果与实测损伤区域的空间重叠质量，
能够反映模型定位准确性平均精度 (mAP) $ \text{mAP=}\dfrac{1}{n}\displaystyle\sum\limits_{i=1}^{n}\text{A}{\text{P}}_{i} $，
$ \mathrm{A}{\mathrm{P}}_{\mathrm{i}} $是第i类别的平均精度表示多类别下精确率-召回率曲线的平均面积，
反映模型多类别识别与定位的综合能力覆盖率 $ \mathrm{Coverage}=\dfrac{\text{CorrectlyIdentifiedArea}}{\text{TotalDamageArea}} $ 正确识别出的损伤区域在总损伤区域中的占比，
体现模型空间识别完整性地理定位误差 $ \text{Error}=\sqrt{{\left({x}_{\text{pred}}-{x}_{\text{true}}\right)}^{2}+{\left({y}_{\text{pred}}-{y}_{\text{true}}\right)}^{2}} $
$ ({x}_{\text{pred}},{y}_{\text{pred}}) $是预测坐标
$ \left({x}_{true},{y}_{true}\right) $是真实坐标预测中心与真实中心的欧氏距离，
衡量模型对损伤目标位置的定位精度推理效率(FPS) $ \text{FPS}=\dfrac{N}{T} $
N为处理的图像(或样本)数量，T为对应的时间衡量模型在给定硬件与输入条件下的推理速度，反映其在
大范围灾区快速评估与应急响应场景中的实际部署效率
下载: 导出CSV
表 5 基于深度学习的遥感影像高价值目标灾害评估技术对比
评估方法 类别 方法描述 典型方法 优点 缺点 基于双时相变化检测的灾害评估方法 密集预测 以每个像素为预测单元输出定位结果和损伤等级 Siam-Unet[30]
BAT[31]
SLgViT[32]
MDA-CD[33]
ChangeMamba[35]①适合对损伤进行精细评估
②模型结构简单①缺乏地物拓扑约束，易产生“椒盐噪声”
②大图处理计算冗余高，依赖后处理对象预测 第一阶段生成对象，第二阶段以对象为单位判断其是否损伤和损伤等级 OCD-BDA[41]
DCA-Det[42]
OoCDNet[43]①具有目标语义一致性
②评估结果统计友好，可解释性强①性能高度依赖第一阶段对象提取准确性
②方法复杂度较高
③标注需要更多人工成本基于多时相序列建模的灾害评估方法 -- 通过时间序列数据建模来捕捉损伤动态变化过程 CNN-STS[47]
TKDS-PtNet[20]①损伤动态变化捕捉能力更强
②降低模型在单一时间点的假阳性概率，提高评估精度①多时相图像间存在配准误差
②多时相计算复杂度高，训练成本高基于多模态遥感数据的灾害评估方法 数据级融合 在输入层对多模态数据
进行整合M-UNet[50] ①结构简单
②保留了原始数据全部信息①难以充分挖掘各模态间的深层关联
②难以针对不同模态进行专门的特征提取和建模特征级融合 首先对各模态数据进行独立编码来提取各自特征，然后在中间特征层进行融合 Attention U-Net[51]
M3ICNet[3]①可设计独立编码器提取各模态特征，灵活性高
②在语义层面对特征进行融合，能捕获跨模态语义关联①模态间特征对齐难度大
②需要更多计算资源和训练时间数据受限下的鲁棒灾害评估方法 -- 利用少量标注或无标注数据完成模型自训练，或者通过借助已有源域数据进行迁移学习 BGPLF[52]
GEM[53]
U-BDD++[54]
STCA[4]
TDA-Net[55]①缓解标签数据匮乏问题
②提升模型的泛化能力
③加快模型部署速度①难以保证模型细粒度识别能力
②域间差异大时效果下降
下载: 导出CSV
表 6 基于深度学习的遥感影像高价值目标灾害评估算法性能对比
评估方法 类别 典型方法 数据集 $ F_{1}^{\text{overall}} $ $ F_{1}^{\text{loc}} $ $ F_{1}^{\text{dam}} $ 基于双时相变化检测的灾害评估方法 密集预测 Siam-Unet[30]
BAT[31]
SLgViT[32]
MDA-CD[33]
ChangeMamba[35]xBD
xBD
Haidi/Changing
xBD
xBD71.7
81.3
90.4/88.6
76.4
81.585.9
88.2
-
86.2
87.465.6
78.4
-
72.1
78.9对象预测 OCD-BDA[41]
DCA-Det[42]
OoCDNet[43]Turkish earthquake dataset
AICD-2012
xBD93.0 (OA)
79.8
71.7-
-
--
-
-基于多时相序列建模的灾害评估方法 -- CNN-STS[47]
TKDS-PtNet[20]WCP
WCP28.3
83.5-
--
-基于多模态遥感数据的灾害评估方法 数据级融合 M-UNet[50] Shuguang dataset 84.7 - - 特征级融合 Attention U-Net[51]
M3ICNet[3]xBD
WBD/EBD50.1
79.3/75.8-
--
-数据受限下的鲁棒灾害评估方法 -- BGPLF[52]
GEM[53]
U-BDD++[54]
STCA[4]
TDA-Net[55]Bright
HC2012
xBD
xBD
xBD74.4(mIoU)
64.5
-
47.6
77.8-
-
58.2
85.0
--
-
63.8
31.5
-注： $ F_{1}^{\text{overall}} $表示目标定位分数$ F_{1}^{\text{loc}} $和损伤等级分类分数$ F_{1}^{\text{dam}} $的加权求和。仅报告$ F_{1}^{\text{overall}} $ 指标的方法表示其指标计算基于定位与分类均正确的目标。个别方法未采用F1，而使用 mIoU 或 OA 进行评价，已在表中以括号形式标注。在数据受限下的鲁棒评估方法中，仍有部分研究采用xBD数据集，这是因为他们引入了无监督或少样本学习策略，以模拟数据受限条件下的应用场景。
下载: 导出CSV
-
[1] 翁星星, 庞超, 许博文, 等. 面向遥感图像解译的增量深度学习[J]. 电子与信息学报, 2024, 46(10): 3979–4001. doi: 10.11999/JEIT240172.WENG Xingxing, PANG Chao, XU Bowen, et al. Incremental deep learning for remote sensing image interpretation[J]. Journal of Electronics & Information Technology, 2024, 46(10): 3979–4001. doi: 10.11999/JEIT240172. [2] 程塨, 王光兴, 韩军伟. 深度学习遥感变化检测综述: 典型算法及发展趋势[J]. 遥感学报, 2025, 29(6): 1587–1597. doi: 10.11834/jrs.20254441.CHENG Gong, WANG Guangxing, and HAN Junwei. Deep learning for change detection in remote sensing: A review and new outlooks[J]. National Remote Sensing Bulletin, 2025, 29(6): 1587–1597. doi: 10.11834/jrs.20254441. [3] ZHANG Haiming, MA Guorui, WANG Di, et al. M3ICNet: A cross-modal resolution preserving building damage detection method with optical and SAR remote sensing imagery and two heterogeneous image disaster datasets[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2025, 221: 224–250. doi: 10.1016/j.isprsjprs.2025.02.004. [4] ZHENG Zhuo, ZHONG Yanfei, ZHANG Liangpei, et al. Towards transferable building damage assessment via unsupervised single-temporal change adaptation[J]. Remote Sensing of Environment, 2024, 315: 114416. doi: 10.1016/j.rse.2024.114416. [5] 刘思琪, 高智, 陈泊安, 等. 基于图网络的遥感地物关系表达与推理的地表异常检测[J]. 电子与信息学报, 2025, 47(6): 1690–1703. doi: 10.11999/JEIT240883.LIU Siqi, GAO Zhi, CHEN Boan, et al. Earth surface anomaly detection using graph neural network-based representation and reasoning of remote sensing geographic object relationships[J]. Journal of Electronics & Information Technology, 2025, 47(6): 1690–1703. doi: 10.11999/JEIT240883. [6] CHEN Guan, LIU Yong, MA Zhangfeng, et al. Assessing extent of building damage following an earthquake: Case study of the 2023 Turkey-Syria doublet[J]. npj Natural Hazards, 2025, 2(1): 51. doi: 10.1038/s44304-025-00101-7. [7] 陈昊, 周光尧, 王乾通, 等. 基于一致性生成对抗的遥感多时相建筑物变化检测数据对生成技术[J]. 电子与信息学报, 2025, 47(3): 825–838. doi: 10.11999/JEIT240720.CHEN Hao, ZHOU Guangyao, WANG Qiantong, et al. Building change detection data generation technology for multi-temporal remote sensing imagery based on consistent generative adversarial[J]. Journal of Electronics & Information Technology, 2025, 47(3): 825–838. doi: 10.11999/JEIT240720. [8] GUPTA R, GOODMAN B, PATEL N, et al. Creating xBD: A dataset for assessing building damage from satellite imagery[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, Long Beach, USA, 2019: 10–17. (查阅网上资料, 未找到本条文献信息, 请确认). [9] UNOSAT and United Nations Institute for Training and Research (UNITAR). UNOSAT official website[OL]. https://www.unitar.org/unosat, 2025. (查阅网上资料,链接与内容不符). [10] Federal Emergency Management Agency. Damage assessment operations manual: A guide to assessing damage and impact[R]. , 2016. (查阅网上资料, 未找到本条文献报告编号信息, 请确认). [11] Federal Emergency Management Agency. Hazus hurricane model user guidance[R]. , 2018. (查阅网上资料, 未找到本条文献报告编号信息, 请确认). [12] GRÜNTHAL G, MUSSON R, SCHWARZ J, et al. EMS-98 (European Macroseismic Scale)[R]. Report of European Seismological Commission, Strasbourg: ESC, 1998. (查阅网上资料, 未找到本条文献信息, 请确认). [13] IWG-SEM. Satellite-based emergency mapping – guidelines for building damage assessment[R]. Version 1.0. Geneva: IWG-SEM, 2017. (查阅网上资料, 未找到本条文献信息, 请确认). [14] ZHAO Zongze, WANG Fenglei, CHEN Shiyu, et al. Deep object segmentation and classification networks for building damage detection using the xBD dataset[J]. International Journal of Digital Earth, 2024, 17(1): 2302577. doi: 10.1080/17538947.2024.2302577. [15] BRAIK A M and KOLIOU M. Automated building damage assessment and large-scale mapping by integrating satellite imagery, GIS, and deep learning[J]. Computer-Aided Civil and Infrastructure Engineering, 2024, 39(15): 2389–2404. doi: 10.1111/mice.13197. [16] LEE C C, KAUR N, MAHDAVI-AMIRI A, et al. Ida-BD: Pre- and post-disaster high-resolution satellite imagery for building damage assessment from hurricane Ida[J/OL]. https://www.designsafe-ci.org/data/browser/public/designsafe.storage.published/PRJ-3563, 2022. [17] RAHNEMOONFAR M, CHOWDHURY T, and MURPHY R. RescueNet: A high resolution UAV semantic segmentation dataset for natural disaster damage assessment[J]. Scientific Data, 2023, 10(1): 913. doi: 10.1038/s41597-023-02799-4. [18] SUN Yao, WANG Yi, and EINEDER M. Post-earthquake SAR-optical dataset for quick damaged-building detection[C]. IEEE International Geoscience and Remote Sensing Symposium, Athens, Greece, 2024: 3787–3790. doi: 10.1109/IGARSS53475.2024.10641601. [19] CHEN Hongruixuan, SONG Jian, DIETRICH O, et al. Bright: A globally distributed multimodal building damage assessment dataset with very-high-resolution for all-weather disaster response[J]. Earth System Science Data, 2025, 17(11): 6217–6253. doi: 10.5194/essd-17-6217-2025. [20] HOU Zhengyang, QU Ying ZHANG Liqiang, et al. War city profiles drawn from satellite images[J]. Nature Cities, 2024, 1(5): 359–369. doi: 10.1038/s44284-024-00060-6. [21] JIANG Wandong, SUN Yuli, LEI Lin, et al. Change detection of multisource remote sensing images: A review[J]. International Journal of Digital Earth, 2024, 17(1): 2398051. doi: 10.1080/17538947.2024.2398051. [22] WANG Lukang, ZHANG Min, GAO Xu, et al. Advances and challenges in deep learning-based change detection for remote sensing images: A review through various learning paradigms[J]. Remote Sensing, 2024, 16(5): 804. doi: 10.3390/rs16050804. [23] GU Jiancheng, XIE Zhengtao, ZHANG Jiandong, et al. Advances in rapid damage identification methods for post-disaster regional buildings based on remote sensing images: A survey[J]. Buildings, 2024, 14(4): 898. doi: 10.3390/buildings14040898. [24] WANG Lili, WU Jidong, YANG Youtian, et al. Deep learning models for hazard-damaged building detection using remote sensing datasets: A comprehensive review[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024, 17: 15301–15318. doi: 10.1109/JSTARS.2024.3449097. [25] MOYA L, GEIß C, HASHIMOTO M, et al. Disaster intensity-based selection of training samples for remote sensing building damage classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 59(10): 8288–8304. doi: 10.1109/TGRS.2020.3046004. [26] BALLINGER O. Open access battle damage detection via pixel-wise T-test on sentinel-1 imagery[J]. Remote Sensing of Environment, 2025, 331: 115025. doi: 10.1016/j.rse.2025.115025. [27] DIETRICH O, PETERS T, GARNOT V S F, et al. An open-source tool for mapping war destruction at scale in Ukraine using sentinel-1 time series[J]. Communications Earth & Environment, 2025, 6(1): 215. doi: 10.1038/s43247-025-02183-7. [28] BRAIK A M, HAN Xu, and KOLIOU M. A framework for resilience analysis and equitable recovery in tornado-impacted communities using agent-based modeling and computer vision-based damage assessment[J]. International Journal of Disaster Risk Reduction, 2025, 121: 105427. doi: 10.1016/j.ijdrr.2025.105427. [29] WANG Yu, LI Yue, and ZHANG Shufeng. Automatic detection of war-destroyed buildings from high-resolution remote sensing images[J]. Remote Sensing, 2025, 17(3): 509. doi: 10.3390/rs17030509. [30] DUNNHOFER M, ANTICO M, SASAZAWA F, et al. Siam-U-Net: Encoder-decoder Siamese network for knee cartilage tracking in ultrasound images[J]. Medical Image Analysis, 2020, 60: 101631. doi: 10.1016/j.media.2019.101631. [31] LU Wen, WEI Lu, and NGUYEN M. Bitemporal attention transformer for building change detection and building damage assessment[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024, 17: 4917–4935. doi: 10.1109/JSTARS.2024.3354310. [32] QIAO Wenfan, SHEN Li, WANG Wei, et al. A weakly supervised bitemporal scene change detection approach for pixel-level building damage assessment using pre- and post-disaster high-resolution remote sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: 5648523. doi: 10.1109/TGRS.2024.3494257. [33] HAN Dongzhe, YANG Guang, LU Wangze, et al. A multi-level damage assessment model based on change detection technology in remote sensing images[J]. Natural Hazards, 2025, 121(6): 7367–7388. doi: 10.1007/s11069-024-07094-y. [34] GU A and DAO T. Mamba: Linear-time sequence modeling with selective state spaces[C]. First Conference on Language Modeling, Philadelphia, USA, 2024. [35] CHEN Hongruixuan, SONG Jian, HAN Chengxi, et al. ChangeMamba: Remote sensing change detection with spatiotemporal state space model[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: 4409720. doi: 10.1109/TGRS.2024.3417253. [36] WIGUNA S, ADRIANO B, MAS E, et al. Evaluation of deep learning models for building damage mapping in emergency response settings[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024, 17: 5651–5667. doi: 10.1109/JSTARS.2024.3367853. [37] REN Shaoqing, HE Kaiming, GIRSHICK R, et al. Faster R-CNN: Towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137–1149. doi: 10.1109/TPAMI.2016.2577031. [38] WANG Ao, CHEN Hui, LIU Lihao, et al. YOLOv10: Real-time end-to-end object detection[C]. Proceedings of the 38th International Conference on Neural Information Processing Systems, Vancouver, Canada, 2024: 3429. [39] HE Kaiming, GKIOXARI G, DOLLÁR P, et al. Mask R-CNN[C]. Proceedings of the 2017 IEEE International Conference on Computer Vision (ICCV), Venice, Italy, 2017: 2980–2988. doi: 10.1109/ICCV.2017.322. [40] CHENG Bowen, MISRA I, SCHWING A G, et al. Masked-attention mask transformer for universal image segmentation[C]. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, USA, 2022: 1280–1289. doi: 10.1109/CVPR52688.2022.00135. [41] XIE Zhengtao, ZHOU Zifan, HE Xinhao, et al. Methodology for object-level change detection in post-earthquake building damage assessment based on remote sensing images: OCD-BDA[J]. Remote Sensing, 2024, 16(22): 4263. doi: 10.3390/rs16224263. [42] ZHANG Lin, HU Xiangyun, ZHANG Mi, et al. Object-level change detection with a dual correlation attention-guided detector[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 177: 147–160. doi: 10.1016/j.isprsjprs.2021.05.002. [43] ZHANG Haiming, ZHANG Yongxian, WANG Di, et al. Damaged building object detection from bitemporal remote sensing imagery: A cross-task integration network and five datasets[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: 5648827. doi: 10.1109/TGRS.2024.3493886. [44] ZHENG Zhuo, ZHONG Yanfei, WANG Junjue, et al. Building damage assessment for rapid disaster response with a deep object-based semantic change detection framework: From natural disasters to man-made disasters[J]. Remote Sensing of Environment, 2021, 265: 112636. doi: 10.1016/j.rse.2021.112636. [45] ZHANG Hongyan, LIN Manhui, YANG Guangyi, et al. ESCNet: An end-to-end superpixel-enhanced change detection network for very-high-resolution remote sensing images[J]. IEEE Transactions on Neural Networks and Learning Systems, 2023, 34(1): 28–42. doi: 10.1109/TNNLS.2021.3089332. [46] ZHAN Tao, GONG Maoguo, JIANG Xiangming, et al. S3Net: Superpixel-guided self-supervised learning network for multitemporal image change detection[J]. IEEE Geoscience and Remote Sensing Letters, 2023, 20: 5002205. doi: 10.1109/LGRS.2023.3300308. [47] MUELLER H, GROEGER A, HERSH J, et al. Monitoring war destruction from space using machine learning[J]. Proceedings of the National Academy of Sciences, 2021, 118(23): e2025400118. doi: 10.1073/pnas.2025400118. [48] LI Jiaxin, HONG Danfeng, GAO Lianru, et al. Deep learning in multimodal remote sensing data fusion: A comprehensive review[J]. International Journal of Applied Earth Observation and Geoinformation, 2022, 112: 102926. doi: 10.1016/j.jag.2022.102926. [49] SAIDI S, IDBRAIM S, KARMOUDE Y, et al. Deep-learning for change detection using multi-modal fusion of remote sensing images: A review[J]. Remote Sensing, 2024, 16(20): 3852. doi: 10.3390/rs16203852. [50] LV Zhiyong, HUANG Haitao, GAO Lipeng, et al. Simple multiscale UNet for change detection with heterogeneous remote sensing images[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19: 2504905. doi: 10.1109/LGRS.2022.3173300. [51] ADRIANO B, YOKOYA N, XIA Junshi, et al. Learning from multimodal and multitemporal earth observation data for building damage mapping[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 175: 132–143. doi: 10.1016/j.isprsjprs.2021.02.016. [52] LI Jiepan, HUANG He, SHENG Yu, et al. Building-guided pseudo-label learning for cross-modal building damage mapping[C]. IGARSS 2025-2025 IEEE International Geoscience and Remote Sensing Symposium, Brisbane, Australia, 2025: 228–232. doi: 10.1109/IGARSS55030.2025.11243835. [53] ZHAO Chunhui, SHEN Yi, SU Nan, et al. Gully erosion monitoring based on semi-supervised semantic segmentation with boundary-guided pseudo-label generation strategy and adaptive loss function[J]. Remote Sensing, 2022, 14(20): 5110. doi: 10.3390/rs14205110. [54] ZHANG Yiyun, WANG Zijian, LUO Yadan, et al. Learning efficient unsupervised satellite image-based building damage detection[C]. 2023 IEEE International Conference on Data Mining (ICDM), Shanghai, China, 2023: 1547–1552. doi: 10.1109/ICDM58522.2023.00206. [55] ZHANG Haiming, WANG Mingchang, ZHANG Yongxian, et al. TDA-Net: A novel transfer deep attention network for rapid response to building damage discovery[J]. Remote Sensing, 2022, 14(15): 3687. doi: 10.3390/rs14153687. -
下载:
图(6) / 表(6)
计量
- 文章访问数: 22
- HTML全文浏览量: 16
- PDF下载量: 0
- 被引次数: 0
